Patients with autosomal recessive protein kinase C δ (PKCδ) deficiency suffer from childhood-onset autoimmunity, including systemic lupus erythematosus. They also suffer from recurrent infections that overlap with those seen in patients with chronic granulomatous disease (CGD), a disease caused by defects of the phagocyte NADPH oxidase and a lack of reactive oxygen species (ROS) production. We studied an international cohort of 17 PKCδ-deficient patients and found that their EBV-B cells and monocyte-derived phagocytes produced only small amounts of ROS and did not phosphorylate p40phox normally after PMA or opsonized Staphylococcus aureus stimulation. Moreover, the patients’ circulating phagocytes displayed abnormally low levels of ROS production and markedly reduced neutrophil extracellular trap formation, altogether suggesting a role for PKCδ in activation of the NADPH oxidase complex. Our findings thus show that patients with PKCδ deficiency have impaired NADPH oxidase activity in various myeloid subsets, which may contribute to their CGD-like infectious phenotype.
Introduction
Protein kinase C δ (PKCδ) is a cytosolic kinase of the novel serine/threonine PKC group (diacylglycerol [DAG]-dependent, Ca2+-independent). It is ubiquitously expressed and is an essential regulator of immune homeostasis and B cell development in humans and mice (Guo et al., 2004; Griner and Kazanietz, 2007). It can be activated in response to various stimuli, leading to individual patterns of PKCδ phosphorylation and, thus, to the differential activation of various downstream targets. PKCδ is involved in regulating apoptosis, proliferation, and cell survival in various cells, including lymphocytes and phagocytes (Salzer et al., 2016; Reyland et al., 1999; Wu-Zhang et al., 2012; Gomel et al., 2007; DeVries-Seimon et al., 2007). It is activated by stimulation with DAG and phospholipids produced by the phospholipase Cγ (PLCγ)–mediated hydrolysis of membrane inositol and phorbol esters (Nishizuka, 1995; Duquesnes et al., 2011). In addition to the conformational changes induced by binding to DAG, PKCδ requires autophosphorylation and phosphorylation by phosphoinositide-dependent kinase-1 (PDK-1) for activation (Durgan et al., 2007). Its enzymatic activity is modulated by members of the Src kinase family, protein tyrosine kinase 2 (PYK2), mechanistic target of rapamycin, and phosphoinositide 3-kinase (PI3K; Duquesnes et al., 2011; Li et al., 1994). PKCδ activity is regulated via its regulatory domain, which contains a pseudosubstrate domain and two conserved domains (C1 and C2-like), with hydrophobic residues within the C1 motif responsible for the binding of PKCδ to DAG and phorbol esters (e.g., PMA; Zhang et al., 1995; Kazanietz et al., 1995). The pseudosubstrate domain, which is located between the C1 and C2-like domains, blocks access to the substrate-binding pocket, keeping PKCδ in an inactive state. The enzymatic activity of PKCδ is dependent on the catalytic domain, composed of the C3 and C4 domains, containing the ATP- and substrate-binding sequences, respectively (Duquesnes et al., 2011; Pappa et al., 1998; Cho, 2001). Several proteins, including STAT1 and -3, nuclear factor of IL-6 (NF-IL6), and ERK1 and -2, have been identified as substrates of PKCδ, implicating this protein in diverse pathways (Salzer et al., 2016; Yang et al., 2019; Nishikawa et al., 1997).
Studies of Prkcd knockout mice have revealed an essential role for PKCδ in B cell homeostasis and tolerance, by demonstrating an accumulation of autoreactive B cells due to defective negative selection in germinal centers and autonomous B cell hyperproliferation (Mecklenbräuker et al., 2002; Limnander et al., 2011). Prkcd-knockout mice thus develop systemic autoimmunity, as shown by the presence of autoantibodies, immune complex–mediated glomerulonephritis, lymphadenopathy, and hepatosplenomegaly, with B cell infiltration in several organs and tissues (Mecklenbräuker et al., 2004; Miyamoto et al., 2002). Bone marrow neutrophils of PKCδ Y155F knock-in mice, lacking the key Y155 phosphorylation site and consequently displaying impaired translocation and activation of PKCδ, have low levels of ROS production and impaired neutrophil extracellular trap (NET) formation (Soroush et al., 2019; Humphries et al., 2008). PKCδ-deficient mice are highly susceptible to Candida albicans and Aspergillus fumigatus, probably due to impaired fungal killing by bone marrow neutrophils and impaired ROS production, as shown for C. albicans (Li et al., 2016). PKCδ also plays a crucial role in the macrophage-mediated phagosomal clearance of Listeria monocytogenes and Mycobacterium tuberculosis via the production of ROS and reactive nitrogen species (Schwegmann et al., 2007). In murine and guinea pig neutrophils, PKCδ is required for the phosphorylation of p40phox, a subunit of the NADPH oxidase complex (Li et al., 2016; Someya et al., 1999). In particular, it mediates phosphorylation of the threonine 154 (T154) residue of p40phox, a key regulatory step in the activation of the NADPH oxidase complex in peripheral neutrophils and B cells, in both mice and humans (Chessa et al., 2010; Grandvaux et al., 2001; Belambri et al., 2018; Matute et al., 2005; Bouin et al., 1998). In human peripheral monocytes, peripheral B cells, and neutrophils, PKCδ mediates phosphorylation of the other phagocyte NADPH oxidase (phox) cytosolic subunits p67phox, p47phox, and p40phox, suggesting a more prominent regulatory function in ROS production (Grandvaux et al., 2001; Zhao et al., 2005; Bey et al., 2004; Brown et al., 2003).
Since 2013, 11 patients from 7 kindreds suffering from autosomal recessive (AR) PKCδ deficiency have been reported (Kuehn et al., 2013; Belot et al., 2013; Kiykim et al., 2015; Salzer et al., 2013; Lei et al., 2018; Nanthapisal et al., 2017; Meyts et al., 2021; Sharifinejad et al., 2020). They manifested with a monogenic form of systemic lupus erythematosus (SLE) occurring before the age of 10 yr. These patients had small numbers of memory B cells (CD19+/CD27+), very large numbers of naive (CD19+/CD27−) and immature (CD19+/CD21−) B cells, and an impaired Ig class switch (Kuehn et al., 2013; Belot et al., 2013). High counts of circulating CD8+ T cells and double-negative (CD4−/CD8−) T cells were observed in some cases, together with slightly low levels of T cell proliferation in vitro (Kuehn et al., 2013; Sharifinejad et al., 2020; Kiykim et al., 2015). Intriguingly, these patients suffer not only from autoimmune disease, but also from recurrent and/or severe infections (Kuehn et al., 2013; Belot et al., 2013; Kiykim et al., 2015; Salzer et al., 2013; Lei et al., 2018; Nanthapisal et al., 2017; Meyts et al., 2021; Sharifinejad et al., 2020). In particular, PKCδ-deficient patients present viral, bacterial, and fungal infections, mostly affecting the lung, gastrointestinal tract, and lymph nodes, reminiscent of those seen in patients with chronic granulomatous disease (CGD). CGD is an inborn error of immunity caused by loss-of-function (LOF) or hypomorphic mutations of any of six genes encoding components of the NADPH oxidase complex (Thomas et al., 2019; Anjani et al., 2019). Like patients with CGD, PKCδ-deficient individuals have been reported to suffer from recurrent, and even invasive, bacterial infections, including in particular infections with Staphylococcus spp. and Pseudomonas spp. (Kuehn et al., 2013; Belot et al., 2013; Kiykim et al., 2015; Salzer et al., 2013; Lei et al., 2018; Nanthapisal et al., 2017; Sharifinejad et al., 2020). The immunosuppressive treatment of SLE is a potential confounding factor, but we hypothesized that these patients might display an intrinsic susceptibility to infections. Based on the role of PKCδ in NADPH oxidase activation, and the overlap between the infections seen in patients with CGD and those with PKCδ deficiency, we tested the hypothesis that AR PKCδ deficiency impairs ROS production in phagocytic cells, thereby underlying susceptibility to CGD-like infectious diseases. We therefore analyzed NADPH oxidase expression and activity in various subsets of myeloid cells, and in EBV-immortalized B cells (EBV-B cells) from all 11 previously reported patients and 6 newly identified patients with AR PKCδ deficiency.
Results
Clinical description of 17 patients with PKCδ deficiency
We studied 17 patients from 10 families originating from Mexico (kindred A), Turkey (kindreds B, D, H, and I), France (kindred C), Pakistan (kindred E), the UK (kindred F), and Iran (kindreds G and J; Table 1 and Fig. 1 A). The 10 families were unrelated, and 6 were known to be consanguineous. 11 patients (P1–P10, P12) from 7 kindreds (A–G) have already been reported to have PKCδ deficiency (Kuehn et al., 2013; Belot et al., 2013; Kiykim et al., 2015; Salzer et al., 2013; Lei et al., 2018; Nanthapisal et al., 2017; Meyts et al., 2021; Sharifinejad et al., 2020). We identified five new patients by whole-exome sequencing (WES). The detailed clinical data of one patient (P11) will be reported elsewhere (unpublished data). Most patients (P1–P15) developed clinical manifestations before the age of 10 yr and displayed hepatosplenomegaly and various features of SLE, including photosensitivity, malar rash, alopecia, and oral ulcers, whereas P16 and P17 displayed no SLE-like symptoms. Furthermore, vasculitis and glomerulonephritis were documented in four (P1, P4, P7, and P8) and autoimmune hemolytic anemia in six (P1, P4, P7, P8, P14, and P15) patients. Most patients displayed B lymphoproliferation and/or high levels of autoreactive antinuclear antibodies (ANAs; P1–P15) or anti–double-stranded DNA (anti-dsDNA) antibodies, as found in nine patients (P3–P9, P12, and P13). In addition to SLE, the patients suffered from recurrent and/or severe infections (Table 1). All patients were vaccinated with Bacillus Calmette-Guérin (BCG) against tuberculosis, but only four patients (P13–P16) developed BCG-related disease. In addition, P13 also suffered from Salmonella bacteremia, recurrent gingivitis, and inguinal lymphadenitis due to Staphylococcus aureus. One patient suffered from severe Legionella pneumonia (P3); one from Streptococcus pneumoniae pneumonia (P12); one from Listeria meningitis, sepsis due to Enterococcus faecium, systemic Candida infection, and recurrent Achromobacter xyloxidans cervical lymphadenitis (P10); and one from sepsis caused by Haemophilus influenzae (P11). P4 died at the age of 13 yr, from septic shock due to Pseudomonas aeruginosa (Table 1). Five patients suffered from viral infections, most caused by herpesviruses, namely EBV (P1), CMV (P2), and varicella-zoster virus (P15), or other viruses, including papillomavirus (P3) and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2; P10). Six patients suffered from recurrent pneumonia (P1, P2, P3, P5, P6, and P12), four from recurrent otitis media (P1, P6, P10, and P12), and six from recurrent gastroenteritis (P2, P6, P11, P12, P14, and P15), although no causal microbe could be identified in any of these cases. Most patients were treated with corticosteroids (P1, P3–P9, and P12–P15), in combination with anti-CD20 antibodies in some cases (P4 and P6–P11), whereas one patient (P2) was successfully treated with hydroxychloroquine alone. Three patients received mycophenolate mofetil (MMF; P4, P5, and P11), two received azathioprine (P7 and P8), and one was treated with sirolimus (P10). Most of the patients have received intravenous immunoglobulin (IVIG; P2, P5–P11, P13, and P15) and/or antibiotics (P2, P6, P12, and 13) as prophylaxis (Table 1).
Biallelic PRKCD variants in 17 individuals from 10 families
Eight disease-causing PRKCD variants in 12 patients have been found: 10 patients (P1–P9 and P12) were found to be homozygous for rare PRKCD variants, including four missense variants (c.1840C>T [p.R614W], c.742G>A [p.G248S], c.1528G>A [p.G510S], and c.1294G>T [p.G432W]), one small insertion (c.1293_1294insA, predicted p.G432Rfs*15), and one essential splice site substitution (c.1352 + 1G>A; Table 1 and Fig. 1 A; Kuehn et al., 2013; Kiykim et al., 2015; Belot et al., 2013; Salzer et al., 2013; Lei et al., 2018; Nanthapisal et al., 2017; Sharifinejad et al., 2020). Two patients, P10 and P11, were compound heterozygous for a nonsense variant (c.571C>T [p.Q191*]) and a variant located at the essential splice acceptor site at the end of intron 9 (c.788-2A>G; Meyts et al., 2021; unpublished data). The newly identified patients, P13–P17, were found to carry previously unknown homozygous PRKCD variants: P13 carried an essential splice site variant (c.571 + 2dup) located in intron 7, P14 and P15 carried a nonsense variant (c.1384C>T [p.Q462*]), and P16 and P17 carried a small homozygous deletion of 1 bp (c.642del) predicted to lead to a premature stop codon (p.N214Kfs*26; Fig. 1, A and B; and Table 1). All variants, in all patients, were confirmed by Sanger sequencing, and when available, familial segregation analysis was performed, confirming that the disease followed an AR pattern of inheritance (Fig. 1 A). Surprisingly, one individual was homozygous for a variant but was asymptomatic at the age of 6 yr (kindred J, II.1, P17). 8 of the 11 variants had already been reported, but we completed the study in silico. 6 of the 11 variants studied were private and were absent from public databases (p.G432W, c.571 + 2dup, c.642del, c.788-2A>G, c.1293_1294insA, and c.1352 + 1G>A), whereas 5 were reported in the heterozygous state in the Genome Aggregation Database (gnomAD v2.1.1; p.Q191*, p.G248S, p.Q462*, p.G510S, and p.R614W), with a minor allele frequency <10−4 (Fig. 1 C). No homozygotes for any of these variants were found in public databases. All missense variants affected sites known to have been conserved over evolution and had combined annotation-dependent depletion (CADD) scores well above the mutation significance cutoff (MSC) of 26.5, whereas the variants predicted to affect splicing had a CADD score close to or slightly below the MSC (Figs. 1 C and S1 A). Only six other nonsynonymous variants of PRKCD were found in the homozygous state in the gnomAD database (v2.1.1) and our in-house cohort (Laboratory of Human Genetics of Infectious Diseases) of >9,500 patients with various infectious diseases. They were missense variants with CADD scores below the MSC and were, therefore, not predicted to be deleterious. These data strongly suggest that the 17 patients studied have AR PKCδ deficiency, consistent with previous reports (Kuehn et al., 2013; Belot et al., 2013; Kiykim et al., 2015; Salzer et al., 2013; Lei et al., 2018; Nanthapisal et al., 2017; Sharifinejad et al., 2020).
PRKCD splice site variants are predicted to encode truncated proteins
Essential splice site variants have been reported for PRKCD, but nothing is known about the consequences of these variants for splicing (Salzer et al., 2013; Meyts et al., 2021). We assessed the impact of all the essential splice site variants identified, by performing topoisomerase-based cloning (TOPO-TA) on full-length cDNA and sequencing 100 transcripts for patients and controls (Fig. S1 B). We detected two aberrantly spliced variants for c.1352 + 1G>A (P6), leading to the skipping of exon 14 (22%; c.1261_1352del; p.D421Vfs*5) or to a 4-bp insertion between exons 14 and 15 (78%; c.1352_1353ins4; p.F452Yfs*6). For c.788-2A>G/p.Q191* (P10 and P11), five transcripts were detected by TOPO-TA cloning: 52% of them corresponded to the p.Q191* nonsense mutation, while 48% were aberrantly spliced, leading to a 38-bp deletion at the start of exon 10 (6%; c.788_825del; p.D263Gfs*14), the skipping of exon 10 (8%; c.788_888del; p.D263Efs*101), or the retention of 124 bp (16%; c.787_788ins124; p.D263Efs*9) or 140 bp (18%; c.787_788ins140; p.D263Vfs*34) of intron 9, respectively. Finally, the c.571 + 2dup (P13) allele encoded two aberrant transcripts, leading to the retention of 25 bp (45%; c.571_572ins25; p.Q191Rfs*1) or 29 bp (55%; c.571_572ins29; p.Q191Rfs*1) of intron 7, resulting in the same predicted truncated protein. Thus, all the detected transcripts were predicted to encode truncated proteins (Fig. S1 B). We also performed TOPO-TA cloning for the missense variant p.G248S (P2), as in silico analysis predicted the creation of a new splice acceptor site in exon 9. However, all the transcripts analyzed corresponded to the patient’s variant, and no aberrantly spliced transcripts were detected. No WT transcripts were detected on analysis of the cDNA obtained from any of the patients. TOPO-TA cloning was not performed for the other variants, as in silico analysis predicted no impact on splicing, and no additional transcripts were detected after gel electrophoresis and full-length PCR.
Seven of 11 PKCδ variants are complete LOF for autophosphorylation when overexpressed
There is strong evidence to suggest that all the reported PRKCD variants are disease causing, but few experimental data are available to decipher the impact of these variants in overexpression systems. We therefore studied the impact of the PRKCD variants by transiently transfecting human embryonic kidney (HEK)293T cells with C-terminally DDK-tagged WT (NM_006254) or mutant PRKCD cDNAs encoding nonsense or missense mutations (p.Q191*, p.G248S, p.G432W, p.Q462*, p.G510S, and p.R614W), cDNAs encoding all PRKCD transcripts identified by TOPO-TA cloning or predicted to cause a frameshift (c.571_572ins25, c.571_572ins29, c.642del, c.787_788ins124, c.787_788ins140, c.788_825del, c.788_888del, c.1293_1294insA, c.1261_1352del, and c.1352_1353ins4), and the cDNA for a previously described LOF mutant (c.1133A>T [p.K378M]), which served as negative control (Belot et al., 2013). All PRKCD mRNAs were produced in similar amounts, as shown by real-time quantitative PCR (RT-qPCR; Fig. 2 A). Protein expression and activity in the basal state were analyzed by Western blotting with antibodies against specific phosphorylation sites of PKCδ known to be required for protein activity (Fig. 2 B; Newton, 2001). The p.G248S, p.G432W, p.G510S, p.R614W, and p.K378M PKCδ variants were expressed at the same molecular weight as the WT protein, whereas transient transfection with cDNAs encoding nonsense and frameshift PRKCD variants led to the expression of truncated proteins of lower molecular weight, consistent with in silico predictions (Figs. 2 B and S1 B). No reinitiation of translation downstream from the premature stop codons was observed, in analyses with an anti-DDK antibody (not depicted). All nonsense and frameshift variants were LOF for the autophosphorylation of PKCδ at the T505 and S643 residues, whereas p.G432W, p.G510S, and p.R614W were hypomorphic (Fig. 2, B and C). The p.G248S variant displayed no impairment of autophosphorylation. Taken together, these results indicate that the nonsense and frameshift mutations strongly impair PKCδ protein expression and abolish PKCδ autophosphorylation, and that all but one of the missense mutations strongly impair autophosphorylation, suggesting that these mutations are responsible for AR PKCδ deficiency.
Altered PKCδ expression and function in the EBV-B cells of eight PKCδ-deficient patients
PKCδ is strongly expressed in human peripheral B lymphocytes and EBV-B cells (Mecklenbräuker et al., 2004; Kuehn et al., 2013; Salzer et al., 2013). We used EBV-B cells to assess the impact of the various mutations on mRNA and protein levels in an endogenous system. We evaluated PRKCD transcript levels in cells from patients and healthy controls by RT-qPCR, with two different probes spanning the junctions of exons 3 and 4 and of exons 17 and 18 (Fig. 3 A). No EBV-B cells were available for P4, P7–P9, P12, or P14–P17. Low levels of PRKCD transcripts were detected in EBV-B cells from P6, P10, P11, and P13; these transcripts carried nonsense or essential splice site mutations and were shown to encode truncated proteins (Figs. 3 A and S1 B). However, for P1–P3 and P5, PRKCD transcript levels were within the range of controls (Fig. 3 A). We further investigated PKCδ protein levels by Western blotting with whole-cell lysates. A total loss of PKCδ protein was observed in the EBV-B cells of P6, P10, P11, and P13 (Fig. 3 B). The essential splice site and nonsense mutations of these patients were predicted to result in truncated proteins, which were detected in overexpression experiments in HEK293T cells, but not in the patients’ cells. However, low (P1, P3, and P5) or high (P2) PKCδ protein levels were detected in EBV-B cells carrying missense mutations (Fig. 3 B). We also assessed the functional consequences of PKCδ deficiency for autophosphorylation and PMA-induced apoptosis, as PKCδ is involved in proapoptotic signaling after phorbol ester stimulation or DNA damage (Fig. 3, C and D; Basu and Pal, 2010). Consistent with our previous overexpression findings, autophosphorylation was abolished (P1, P6, P10, P11, and P13) or impaired (P3 and P5) in the patients’ EBV-B cells, except for those of P2, which displayed normal autophosphorylation (Fig. 3 C). However, the EBV-B cells of P2 displayed impaired myristoylated alanine-rich C-kinase substrate (MARCKS) phosphorylation in response to PMA stimulation (Fig. S2). The EBV-B cells of all the patients (including P2) were resistant to PMA-induced cell death but remained as sensitive as the EBV-B cells of healthy controls to alternative apoptosis inducers, such as FAS (Fig. 3 D; Kuehn et al., 2013). Overall, these findings suggest that PRKCD mRNAs carrying nonsense or frameshift mutations undergo nonsense-mediated mRNA decay, resulting in a loss of expression of the PKCδ protein, whereas at least two of the missense mutations are associated with the production of smaller amounts of PKCδ protein. Strikingly, the EBV-B cells of all patients tested presented a defect of PMA-induced apoptosis, indicating a loss of PKCδ function.
Impaired NADPH oxidase activity and complementation in the patients’ EBV-B cells
Members of the PKC family are known to be involved in NADPH oxidase activity in murine phagocytes, and PKCδ, in particular, has been shown to regulate the activation of NADPH oxidase by phosphorylating p40phox (Li et al., 2016; Someya et al., 1999). However, the role of PKCδ in human phagocytes and EBV-B cells has been less studied (Li et al., 2016; Szilagyi et al., 2015). We therefore evaluated NADPH oxidase activity in EBV-B cells from patients with PKCδ deficiency, comparing the results with those for healthy controls, and cells from patients with CGD (X-linked recessive complete gp91phox deficiency and AR p40phox deficiency). The PMA-induced production of intracellular superoxide (O2−), measured in a luminol chemiluminescence assay, was strongly impaired in the EBV-B cells of all PKCδ-deficient patients tested relative to healthy controls, whereas the EBV-B cells of p40phox- and gp91phox-deficient patients had impaired and abolished O2− production, respectively (Fig. 4 A, left). Similarly, a strong decrease in O2− production by EBV-B cells was observed for all PKCδ-deficient patients tested following stimulation with heat-killed S. aureus cells (Pansorbin), for receptor-mediated NADPH oxidase activation (Fig. 4 A, right). In addition, analyses of Amplex Red oxidation after PMA stimulation showed that much less hydrogen peroxide (H2O2) was released by the EBV-B cells of all PKCδ-deficient patients tested than by the EBV-B cells of healthy controls, with the amount of H2O2 released by the patients’ cells being similar to that released by EBV-B cells from patients with p40phox deficiency (Fig. 4 B). H2O2 production in response to PMA was restored in the EBV-B cells of all patients by retroviral transduction with the WT PRKCD cDNA, as shown by Amplex Red oxidation (Figs. 4 C and S3). We also evaluated the impact of all the PRKCD alleles identified on ROS production by EBV-B cells, by transducing EBV-B cells from P13 (PRKCD−/−) with WT and all mutant-PRKCD cDNAs. In PRKCD−/− EBV-B cells, H2O2 production was restored only by the overexpression of the PRKCD WT allele. H2O2 production was not increased by transduction with any of the mutant alleles (Fig. 4 D). These findings demonstrate that PKCδ plays a nonredundant role in the activation of the NADPH oxidase in EBV-B cells.
Interaction of PKCδ with p40phox/p47phox and impaired p40phox phosphorylation in the patients’ EBV-B cells
We investigated the mechanism underlying the impaired NADPH oxidase activity, by first evaluating the expression of all known NADPH oxidase subunits in the patients’ EBV-B cells. Western blotting showed that all subunits of the NADPH oxidase (gp91phox, p67phox, p47phox, p40phox, p22phox, and EROS) were produced in normal amounts in EBV-B cells from all the PKCδ-deficient patients studied (Fig. 5 A). Flow cytometry confirmed the normal levels of gp91phox, p67phox, p47phox, p22phox, and cytochrome b558, the membrane complex formed by gp91phox and p22phox (Fig. S4 A). We then investigated possible interactions between PKCδ and all known subunits of the NADPH oxidase, by performing coimmunoprecipitation experiments in an overexpression system based on HEK293T cells. Immunoprecipitation and immunoblotting confirmed an interaction between p40phox and PKCδ and between p47phox and PKCδ (Fig. 5 B and Fig. S4, B and C). However, PKCδ did not interact with the other cytosolic (p67phox and RAC2) or membrane-bound (gp91phox, p22phox, and EROS) subunits (Fig. 5, C–E; and Fig. S4, D and E). Overexpression of PKCδ with p40phox or p47phox in HEK293T cells showed that PKCδ was able to phosphorylate both proteins when coexpressed (Fig. 5, F and G). Interestingly, the level of p40phox phosphorylation at the T154 residue was much lower in the EBV-B cells of the PKCδ-deficient patients than in those of the controls after PMA and S. aureus stimulation, whereas the phosphorylation of p47phox at two known PKCδ target residues (S304 and S315) remained similar to that in control EBV-B cells (Fig. 6, A–D; Fontayne et al., 2002). Phosphorylation at other known PKCδ-target sites of p47phox (S320, S328, and S379) could not be evaluated due to a lack of specific antibodies. The transduction of EBV-B cells with PRKCD WT cDNA restored p40phox phosphorylation, for all the patients studied (Fig. 6 E). In conclusion, the EBV-B cells of patients with PKCδ deficiency have impaired ROS production, associated with lower levels of phosphorylation of the cytosolic NADPH oxidase subunit p40phox by PKCδ.
Impact of PKCδ deficiency on NADPH oxidase activity in the patients’ phagocytic cells
Because the EBV-B cells of PKCδ-deficient patients had very low levels of ROS production, we investigated NADPH oxidase activity in response to various stimuli in the primary cells of the patients. In particular, we assessed dihydrorhodamine 123 (DHR) oxidation, which detects intracellular H2O2 production, in neutrophils and monocytes following stimulation with PMA or serum-opsonized Escherichia coli. The patients’ neutrophils and monocytes displayed impaired but not abolished DHR oxidation in response to both stimuli, like the cells of patients with inherited p40phox deficiency (Fig. 7, A and B; and Fig. S5, A and B). In addition, neutrophils and monocytes of P3 and P13 showed low or abolished PKCδ expression, respectively (Fig. S5 C). We also evaluated the formation of NETs by neutrophils from three patients (P3, P16, and P17) from two unrelated kindreds following stimulation with PMA. NET formation was strongly impaired in the patients’ cells, whereas it was abolished in cells from a gp91phox-deficient patient (Figs. 7 C and S5 D). We also explored NADPH oxidase activity in monocyte-derived macrophages (MDMs). MDMs from all the patients tested displayed impaired H2O2 release following stimulation with IFN-γ, IFN-γ plus PMA, or PMA alone, relative to healthy controls or patients with p40phox deficiency (Figs. 7 D and S5 E). Similarly, the NADPH oxidase activity of monocyte-derived dendritic cells (MDDCs), evaluated by assessing H2O2 release after stimulation with LPS, LPS plus PMA, or PMA, was weaker than that of cells from healthy controls and p40phox-deficient patients (Fig. 7 E). As observed for EBV-B cells, the patients’ MDMs displayed reduced or abolished PKCδ expression as well as impaired p40phox phosphorylation at the T154 residue, but normal p47phox phosphorylation at two other residues (S304 and S315) tested (Figs. 7 F and S5 F). The primary phagocytes of heterozygous relatives of PKCδ-deficient patients displayed no defects of ROS production (not depicted). Overall, these data suggest that PKCδ is essential for NADPH oxidase activation and NET formation in primary phagocytic cells, possibly through the phosphorylation of p40phox.
Discussion
We report an impairment of NADPH oxidase activity in 17 patients from 10 unrelated families with AR PKCδ deficiency. ROS production was severely impaired in the EBV-B cells, primary phagocytes (neutrophils and monocytes), MDMs, and MDDCs of all patients studied (Table 2). The transduction of the patients’ EBV-B cells with WT PRKCD cDNA restored not only ROS production, but also the phosphorylation of p40phox on the T154 residue after PMA stimulation, strongly suggesting a role for PKCδ in activation of the NADPH oxidase complex. The phosphorylation of the T154 residue of p40phox is a key regulatory step in neutrophils and B lymphocytes, leading to assembly of the soluble cytosolic NADPH oxidase subunits with the membrane cytochrome b558 (Someya et al., 1999; Chessa et al., 2010; Grandvaux et al., 2001; van de Geer et al., 2018; Matute et al., 2005). Indeed, the MDMs of the PKCδ-deficient patients displayed impaired p40phox phosphorylation and very low levels of ROS production. However, the MDMs of p40phox-deficient patients displayed normal ROS production, indicating that these cells are not dependent on p40phox, and therefore on p40phox phosphorylation, for PMA-induced ROS production. These findings suggest the requirement of another PKCδ substrate for normal ROS production.
Coexpression and coimmunoprecipitation experiments showed that PKCδ interacted not only with p40phox, but also with p47phox, and was able to phosphorylate both proteins. Like p40phox, p47phox must undergo phosphorylation at various residues for translocation and binding to p22phox (El-Benna et al., 2009). Interestingly, we detected no abnormal phosphorylation of p47phox on the S304 and S315 residues in EBV-B cells or MDMs from PKCδ-deficient patients. As several PKC isoforms have been reported to mediate p47phox phosphorylation in human neutrophils, this suggests that PKCδ is redundant for p47phox phosphorylation at these residues (Bey et al., 2004; Fontayne et al., 2002; Segal et al., 2000). However, p47phox phosphorylation at S379 appears to be the most relevant for its activation and was shown to be a PKCδ target site (Fontayne et al., 2002; Meijles et al., 2014). Unfortunately, no antibodies are yet available for the evaluation of p47phox phosphorylation at S379, and further studies are required to evaluate the exact contribution of PKCδ to p47phox phosphorylation at various sites.
In all subsets of primary (neutrophils and monocytes) or monocyte-derived (MDMs and MDDCs) phagocytic cells tested after stimulation with PMA or E. coli, we observed no difference in ROS production or p40phox phosphorylation between patients with and without corticosteroid treatment. In addition to the low level of ROS production in neutrophils and monocytes after stimulation with PMA or E. coli, we observed a strong impairment of NET formation in PKCδ-deficient neutrophils, potentially contributing further to the impairment of bactericidal and fungicidal activity. In addition, a previous study reported a neutrophil killing defect in response to E. coli and C. albicans in PKCδ-deficient neutrophils from three of our patients (P3, P5, and P6; Szilagyi et al., 2015). These data are consistent with those obtained for Pkcδ-deficient mice, which display poor ROS production by bone marrow neutrophils, impaired NET formation, and impaired killing of C. albicans (Soroush et al., 2019; Li et al., 2016). Further studies are required to determine the contributions of PKCδ to the assembly and activity of the NADPH oxidase and pathogen killing. In this respect, patient-derived induced pluripotent stem cells could be used to generate various subsets of myeloid cells. A humanized mouse model could also be used to dissect the development of different subsets of phagocytic cells and their contribution to ROS production (Neehus et al., 2018; Sontag et al., 2017; Lachmann et al., 2015; Evren et al., 2020). Overall, our data suggest that PKCδ is important for host innate immunity through its role in NADPH oxidase activation.
Our studies of NADPH oxidase activity in PKCδ-deficient cells revealed a cellular phenotype related to, but different from, that seen in patients with classic CGD, more closely resembling that seen in patients with p40phox deficiency (van de Geer et al., 2018; Wright et al., 2019; Matute et al., 2009). The defect of ROS production in PKCδ-deficient patients probably at least partly accounts for the recurrent and/or severe infections observed in these patients. The patients studied had recurrent bacterial infections, mostly affecting the lungs, lymph nodes, and gastrointestinal tract (Kuehn et al., 2013; Belot et al., 2013; Kiykim et al., 2015; Salzer et al., 2013; Lei et al., 2018; Nanthapisal et al., 2017; Meyts et al., 2021; Sharifinejad et al., 2020; unpublished data). Two patients had invasive infections, one (P10, who survived infection) with Candida and another (P4, who died) with P. aeruginosa. Recurrent pneumonia and infections with S. aureus are also frequently seen in patients with CGD (Anjani et al., 2019; Winkelstein et al., 2000; de Oliveira-Junior et al., 2015; Song et al., 2011). Nevertheless, patients with classic CGD typically suffer from invasive, life-threatening, recurrent infections with specific bacteria and fungi, such as Burkholderia cepacia, Staphylococcus spp., and Aspergillus spp., due to defective ROS production in all phagocytic cells (de Oliveira-Junior et al., 2015; Marciano et al., 2015; Winkelstein et al., 2000).
The infectious phenotype observed in our cohort of PKCδ-deficient patients was milder than that observed in classic CGD. Like patients with CGD or X-linked recessive Mendelian susceptibility to mycobacterial disease (MSMD) due to hypomorphic mutations of CYBB, PKCδ-deficient patients can suffer from BCG disease, with incomplete penetrance (Conti et al., 2016; Norouzi et al., 2012; Bustamante et al., 2011). However, whereas patients with classic CGD or p40phox deficiency do not seem to have any particular predisposition to severe viral illnesses, five PKCδ-deficient patients (P1, P2, P3, P10, and P15) presented with EBV, CMV, papillomavirus, SARS-CoV-2, and varicella-zoster virus infections, respectively (Meyts et al., 2021; Kuehn et al., 2013; Omarjee et al., 2019). Furthermore, none of the reported p40phox-deficient patients suffered from any severe invasive bacterial or fungal infections, whereas PKCδ-deficient patients did not present with any of the cutaneous infections typically seen in patients with p40phox deficiency (Li et al., 2016; van de Geer et al., 2018; Wright et al., 2019; Matute et al., 2009). Interestingly, p40phox-deficient patients, who have low to normal levels of ROS production in neutrophils and monocytes following phorbol ester stimulation, have no detectable impairment of ROS production by MDMs and MDDCs (van de Geer et al., 2018; Wright et al., 2019; Matute et al., 2009). Unlike patients with classic CGD, PKCδ-deficient patients displayed residual ROS production by circulating and monocyte-derived phagocytes, possibly accounting for the milder clinical phenotype of these patients than of patients with CGD (van de Geer et al., 2018; Matute et al., 2009; Song et al., 2011). The differences in clinical presentation between patients with PKCδ deficiency, p40phox deficiency, and classic CGD can be also explained by the mechanisms proposed here, according to which the PKCδ-dependent phosphorylation not only of p40phox, but also of p47phox, is required for ROS production. Finally, the greater susceptibility to viral illnesses observed in PKCδ-deficient patients may result from low levels of NK cell cytolytic activity, as previously shown (Kuehn et al., 2013; Kiykim et al., 2015).
One of the limitations of our study is the small size of the cohort of PKCδ-deficient patients followed over a short period. Studies of a larger number of patients are required to improve comparisons of the immunological and clinical features of PKCδ deficiency, p40phox deficiency, and classic and variant CGD (Royer-Pokora et al., 1986; Dinauer et al., 1987). Nevertheless, our data indicate that PRKCD mutations underlie a new, “syndromic form” of CGD, characterized by susceptibility to some, but not all, of the infections typical of CGD, associated with early-onset SLE. Indeed, the most striking phenotype observed in PKCδ-deficient patients is the early onset of SLE, which was diagnosed in 13 patients of this cohort (P1–P13), with two others presenting signs of autoimmunity (P14 and P15) and two displaying no autoimmune phenotype at the ages of 3 (P16) and 5 (P17) yr (Kuehn et al., 2013; Belot et al., 2013; Kiykim et al., 2015; Salzer et al., 2013; Lei et al., 2018; Nanthapisal et al., 2017; Meyts et al., 2021; Sharifinejad et al., 2020; unpublished data). Several monogenic disorders underlying SLE have been reported, affecting the complement system, type I IFN immunity, nucleic acid metabolism, or B cell development checkpoints (Costa-Reis and Sullivan, 2017; Rahman and Isenberg, 2008; Omarjee et al., 2019). B cell development is tightly regulated to prevent self-reactivity, as B cell receptor diversification inevitably results in some B cell receptors recognizing self-antigens (Meffre and Wardemann, 2008). Autoreactive B cells are eliminated through multiple steps during B cell development. PKCδ plays a key role in regulating B cell survival, differentiation, and apoptosis (Guo et al., 2004; Griner and Kazanietz, 2007). As previously shown in mice, human PKCδ deficiency may underlie autoimmunity due to defective proapoptotic signaling during B cell development (Mecklenbräuker et al., 2002, 2004; Limnander et al., 2014).
Intriguingly, some SLE manifestations are also observed in some patients with p40phox deficiency or classic CGD (van de Geer et al., 2018; Xie et al., 2016; Parvaneh et al., 2014). Despite the overlap between the cellular phenotypes of these conditions and that of PKCδ deficiency in terms of ROS production, the SLE-like manifestations in these other conditions may be unrelated to the role of PKCδ in B cell checkpoints. SLE is often associated with an increase in type I IFN activity and a dysregulation of type I IFN metabolism (Crow, 2011; Picard and Belot, 2017; Rönnblom et al., 1990). In the absence of NADPH oxidase activity, the mitochondrial ROS-dependent NET-osis of low-density granulocytes leads to the externalization of proinflammatory oxidized mitochondrial DNA and the subsequent activation of type I IFN synthesis (Lood et al., 2016). Enhanced type I IFN activity may, therefore, contribute to the SLE-like phenotype observed in patients with impaired NADPH oxidase activity. Overall, the autoimmunity observed in PKCδ-deficient patients is probably mostly due to the defective selection of autoreactive B cells, although defective ROS production may also be involved.
Materials and methods
Clinical information for the newly identified patients
P13 (kindred H, II.1) was born in 2007 to nonconsanguineous parents of Turkish origin. She was vaccinated with BCG, on the left shoulder, at the age of 2 mo and developed BCG-itis 3 mo later. An excisional biopsy of the left axillary lymph node was performed, and M. tuberculosis PCR and culture were negative. At the age of 11 mo, P13 presented with fever and left axillary, retro-auricular, and cervical lymphadenopathies. Salmonella sp. was documented in blood cultures and treated with ceftriaxone for 21 d. The axillary lymph node was removed; histological examination revealed necrotizing granulomatous inflammation, and Mycobacterium bovis–BCG was identified by culture. Serological tests for HIV were negative. The patient was treated with triple antimycobacterial therapy and was diagnosed with MSMD. Whole-blood activation tests showed that the production of IFN-γ and IL-12p40 in response to stimulation with BCG + IL-12 and BCG + IFN-γ was normal. At 18 mo of age, P13 presented with a skin rash, joint pain, and high ANA levels. Drug-induced lupus erythematosus was suspected on clinical grounds, and isoniazid was discontinued. However, physical examination showed a worsening of her condition, with the appearance of lupus-like symptoms (arthralgia, morning stiffness, and limitation of movement). At the age of 24 mo, antimycobacterial treatment was discontinued. Leukocyte, absolute neutrophil, and absolute lymphocyte counts were normal. Serum IgG levels were low, whereas serum IgA, IgM, and IgE levels were within normal ranges. The immunophenotyping of B cell subsets showed predominantly high naive B cell and low memory B cell counts, with markedly high levels of CD21low and transitional B cells. In summary, this patient received multiagent antituberculosis treatment for 33 mo, without relapse or the use of other antimycobacterial treatments. The patient has had recurrent gingivitis since the age of 6 yr and, at the age of 7 yr, presented with inguinal lymphadenitis, leading to the detection of S. aureus by culture. The patient is now 14 yr old and remains free from mycobacterial and other infectious diseases. However, SLE and the high levels of ANA persist.
P14 and P15 (kindred I, II.1 and II.2) were born in 2010 and 2013, respectively, to nonconsanguineous parents of Turkish origin. Both were vaccinated with BCG at 1 mo of age and developed BCG-itis 6 mo later. P14 presented with fever, diarrhea, hepatosplenomegaly, failure to thrive, and left axillary lymphadenopathy with purulent discharge. Lymph node biopsy showed granulomatous inflammation with no microbial cause identified upon culture. Similarly, P15 developed left axillary lymphadenopathy and hepatosplenomegaly after BCG vaccination. Both patients were treated with isoniazid and rifampicin for 4 (P14) and 5 (P15) yr; P15 also received IVIG treatment. Prophylaxis was not indicated in P14. Both patients had high levels of ANA and developed autoimmune anemia, with no other lupus-like symptoms. Both patients are currently well.
P16 and P17 (kindred J, II.2 and II.1) were born in 2017 and 2014, respectively, to consanguineous parents of Iranian origin. Both were vaccinated with BCG at birth, but only P16 developed BCG-itis; her brother (P17) remained asymptomatic. P16 presented at 14 mo of age with bilateral anterior and posterior cervical adenopathies and mild hepatosplenomegaly. Lymph node biopsy showed chronic granulomatous inflammation with caseating necrosis. No microbe was identified, but a purified protein derivative skin test was positive. The patient received triple antimycobacterial therapy (isoniazid, rifampin, and azithromycin) for 6 mo until the adenopathies disappeared. Her serum IgG, IgA, and IgM levels were within the normal range, but her antibody responses to tetanus and diphtheria vaccines were nonprotective (anti-tetanus antibodies, 0.04 mg/dl; anti-diphtheria antibodies, 0.04 mg/dl [normal >0.1 mg/dl]). P16 and P17 both tested negative for autoantibodies and have not yet developed any lupus-like symptoms. Both patients are well, without prophylaxis.
Blood collection
Blood samples were collected from all patients and relatives, at the local institutions, after written informed consent had been obtained. The study was approved by the institutional review boards of the Institut National de la Santé et de la Recherche Médicale (INSERM), Rockefeller University, and Necker Hospital for Sick Children. All these procedures were conducted in accordance with the 1975 Declaration of Helsinki, as revised in 2013.
Molecular genetics
For P13, genomic DNA (gDNA) was isolated from whole blood with the iPrep PureLink gDNA Blood Kit and iPrep Instruments (Life Technologies, Thermo Fisher Scientific). Exome capture was performed with the SureSelect Human All Exon 71 Mb kit (Agilent Technologies), with 3 µg gDNA. Single-end sequencing was performed on an Illumina Genome Analyzer IIx (Illumina). For P14 and P15, gDNA was obtained from the patients’ peripheral blood samples with the QIAamp DNA Blood Mini QIAcube Kit (Qiagen). The library was constructed with a 200-amplitude primary immunodeficiency panel kit from Sistemas Genomicos. Samples and index reading primers were loaded into a 300-cycle MiSeq Reagent Kit v2 300 cartridge (Illumina), and parallel sequencing was performed in MiSeq v2 Flowcell (Illumina). The variants of all the other patients were identified by WES (P1, P2, P3, P6, P7, P12, P16, and P17), Sanger sequencing (P4, P5, P8, and P9), or focused WES with a primary immunodeficiency subpanel for next-generation sequencing (P10–P11), as previously described (Kuehn et al., 2013; Kiykim et al., 2015; Belot et al., 2013; Salzer et al., 2013; Lei et al., 2018; Nanthapisal et al., 2017; Sharifinejad et al., 2020). All mutations and familial segregation patterns were confirmed by amplifying the flanking regions of the variants obtained with specific primers (Table S1). PCR products were analyzed by electrophoresis in 1% agarose gels, sequenced with the Big Dye Terminator v3.1 cycle sequencing kit (Applied Biosystems), and analyzed on an ABI Prism 3700 (Applied Biosystems). The patients reported here did not consent to the deposition of their genomic data.
Expression of mutated PKCδ proteins in transfected HEK293T cells
Site-directed mutagenesis with specific primers (Table S2) was performed to generate the PRKCD variants in the pCMV6-PRKCD-WT-DDK plasmid (#RC221652; OriGene). HEK293T cells were transfected with 1 µg of the plasmid in the presence of X-tremeGENE 9 DNA transfection reagent (Merck) for 24 h. Protein extraction and Western blotting were performed as described below.
TOPO-TA cloning
RNA was extracted from the EBV-B cells of P2, P6, P10, P11, P13, and a healthy control with the Quick-RNA Microprep Kit (Zymo Research), and RNA was transcribed to generate cDNA with the High-Capacity RNA-to-cDNA Kit (Thermo Fisher Scientific). Full-length PRKCD cDNA was amplified by PCR with the following primers: 5′-CTATGAGCTGGGCTCCCTG-3′ and 5′-CCAGGAGGTGCTCGAATTTG-3′. The purified PCR products were inserted into a pCR4-TOPO vector (Thermo Fisher Scientific) and used to transform competent NEB 10-β E. coli cells (New England Biolabs). We picked 100 single colonies per reaction, and the inserted cDNA was amplified with M13 forward and reverse primers.
RT-qPCR
Gene expression was analyzed by RT-qPCR with the TaqMan Universal PCR Master Mix (Thermo Fisher Scientific) and TaqMan probes for PRKCD (Hs01090047_m1 and Hs01090051_m1) and GUSB (4326320E; all from Thermo Fisher Scientific). Gene expression was normalized against GUSB as an endogenous control for each sample and then according to the mean value for controls in one experiment.
Western blot analysis
EBV-B cells were treated with 400 ng/ml PMA (Sigma-Aldrich) or 2 mg/ml Pansorbin (Sigma-Aldrich) in HBSS for 30 min or were left untreated, and total protein extracts were prepared with insoluble protein buffer (20 mM Tris-HCl, pH 8, 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100, supplemented with 1 mM 4-[2-aminoethyl] benzene sulfonyl fluoride hydrochloride, 20 ng/ml chymostatin, 10 nM leupeptin, and 2 mM PMSF; all from Sigma-Aldrich). Protein extracts (30 µg protein per lane) were resolved by electrophoresis in a 10% Criterion TGX precast gel, and the bands obtained were transferred onto a nitrocellulose membrane by standard transfer methods, in a Transblot turbo system (Bio-Rad). Membranes were probed by overnight incubation with the appropriate primary antibodies (Table S3) at 4°C. After incubation with the corresponding secondary antibody (goat anti-rabbit IgG (H+L)-HRP and goat anti-mouse IgG (H+L)-HRP; Bio-Rad), proteins were detected by chemiluminescence. For the quantification of PKCδ protein, membranes were probed with anti-PKCδ and anti-GAPDH antibodies, and binding was visualized with the Odyssey CLx Imaging system (LI-COR) after incubation with IRDye 800CW goat anti-mouse IgG (H+L) and IRDye 680RD goat anti-rabbit IgG (H+L; both LI-COR) secondary antibodies.
Flow cytometry
EBV-B cells were incubated with LIVE/DEAD Fixable Aqua Dead Cell Stain (Thermo Fisher Scientific) for 10 min at 37°C. Cells were fixed and permeabilized in intracellular fixation and permeabilization buffer (eBioscience, Thermo Fisher Scientific) for 30 min. Intracellular staining was performed for 30 min at 4°C with the appropriate antibodies (Table S3). Cells were analyzed by flow cytometry with a Gallios FACS Analyzer.
Coimmunoprecipitation
Coimmunoprecipitation was performed as previously described (Béziat et al., 2018). Briefly, HEK293T cells were cotransfected with constructs encoding Myc-tagged, DDK-tagged, or V5-tagged proteins or the empty pCMV6 or pCDNA3.1 plasmid in the presence of X-tremeGENE 9 DNA transfection reagent. After 24 h, cells were lysed with lysis buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 2 mM EDTA, and 0.5% Triton X-100) supplemented with aprotinin (10 µg/ml; Sigma Aldrich), PMSF (1 mM), and leupeptin (10 µg/ml) for 30 min at 4°C and centrifuged for 5 min at 16,000 ×g. The supernatant was collected for immunoprecipitation. Immunoprecipitation was performed with anti-V5 (#R96025; Thermo Fisher Scientific) or anti-DDK (#OTI4C5; OriGene) antibodies and agarose-A/G beads (Santa Cruz Biotechnology), after a background clearing step with a normal mouse IgG antibody (#sc-2025; Santa Cruz Biotechnology) and agarose-A/G beads. Total and immunoprecipitated proteins were then analyzed by Western blotting, as described above.
Apoptosis measurement
We used 1 × 105 EBV-B cells to inoculate RPMI 1640 containing 10% FCS in the presence or absence of PMA (100 ng/ml) or APO-1-1 (1 µg/ml; Enzo Life Science), in triplicate, in 96-well plates. The cells were cultured for 24 h at 37°C. Cells were stained with LIVE/DEAD Fixable Aqua Dead Cell Stain, and live cells were counted by flow cytometry with constant time acquisition on a Gallios FACS Analyzer. The percentage of live cells was calculated as previously described (Kuehn et al., 2013).
Retroviral transduction of EBV-B cells
Retrovirus-mediated transduction was performed as previously described (Martínez-Barricarte et al., 2016). In brief, the coding sequences of the PRKCD WT and PRKCD mutant alleles were inserted into the pLZRS-IRES-ΔNGFR plasmid. Retroviruses were produced by transfecting Phoenix A cells with the newly generated plasmids, and EBV-B cells from PKCδ-deficient patients were transduced with the viral concentrate obtained. The transduced cells were subjected to magnetic activated cell sorting purification with anti-CD271 MicroBeads (Miltenyi Biotec), and purified cells were used in subsequent experiments. Transduction was confirmed by flow cytometry with anti-CD271-PE antibody (BD Bioscience).
Differentiation of MDMs and MDDCs
MDMs and MDDCs were differentiated as previously described (van de Geer et al., 2018; Bustamante et al., 2011; Conti et al., 2015). In brief, CD14+ cells were isolated from peripheral blood mononuclear cells by positive selection with anti-CD14 MicroBeads (Miltenyi Biotec). For MDM differentiation, cells were cultured in RPMI 1640 containing 10% FCS and M-CSF (50 ng/ml). On day 7, IL-4 (50 ng/ml) was added, and the cells were incubated for a further 7 d for the completion of differentiation. MDDCs were obtained by incubation in the presence of GM-CSF (50 ng/ml) and IL-13 (20 ng/ml; all R&D Systems) for 7 d.
NADPH oxidase functional assays
The intracellular production of O2− in EBV-B cells was assessed by chemiluminescence (Sigma-Aldrich; van de Geer et al., 2018). We stimulated 5 × 105 cells with PMA (400 ng/ml) or heat-killed S. aureus cells (Pansorbin, 2 mg/ml; Sigma-Aldrich) and light production was monitored every 5 min with a Victor X4 plate reader. NADPH oxidase activity was assessed in EBV-B cells, MDMs, and MDDCs by measuring extracellular H2O2 release with an Amplex Red Kit (Thermo Fisher Scientific), as previously described (van de Geer et al., 2018; Bustamante et al., 2011; Conti et al., 2015). In brief, 3 × 104 cells were stimulated with PMA (400 ng/ml), and H2O2 release was quantified with a Victor X4 plate reader (PerkinElmer). MDMs or MDDCs were cultured for 16 to 18 h before the experiment, in the presence of 5 × 103 IU/ml IFN-γ (Imukin, Boehringer Ingelheim) or 1 µg/ml LPS (Salmonella minnesota; Sigma-Aldrich), respectively. ROS production by neutrophils and monocytes was quantified with the Burstest (Phagoburst kit; BD) containing PMA and opsonized E. coli bacteria. The assay was performed according to the manufacturer’s instructions, and the amount of rhodamine 123 (mean fluorescence intensity) was assessed by flow cytometry with a Gallios FACS Analyzer. DHR oxidation is expressed as a percentage of the rhodamine 123 produced by the corresponding travel controls.
NET formation
Neutrophils were isolated from whole blood with density gradient centrifugation and Polymorphprep (Progen). We used 1 × 105 neutrophils to seed 8-well microslides (Ibidi), which were then incubated at 37°C for 15 min. PMA (100 ng/ml) was then added, and the cells were incubated for 3 h at 37°C. The cells were fixed by incubation in 4% paraformaldehyde for 15 min, blocked and permeabilized by incubation with 5% BSA in PBS-Tween for 1 h, and incubated overnight with an anti-myeloperoxidase antibody (ab45977; Abcam). Secondary staining was performed by incubation for 1 h with a goat anti-rabbit Alexa Fluor 488 antibody (Thermo Fisher Scientific). Cells were mounted in ProLong Gold antifade reagent containing DAPI (Thermo Fisher Scientific), and images were acquired at the Necker Institute Imaging Facility with a Leica SP8 gSTED confocal microscope (Leica). Percentage NET formation was quantified by counting cells with decondensed nuclei and extracellular DAPI-positive fibers in at least five image fields selected at random.
Statistics
A linear mixed model was used to analyze the data presented in Fig. 3 D, to take repeated measurements into account. The null hypothesis was that the distribution of EBV-B cell survival rates after PMA or APO stimulation was similar in patients and controls. The null hypothesis was rejected if the P value was <0.05.
Statistical analysis and visualization of results for Fig. 7, A–E, were performed using GraphPad Prism 8.4.3. For Fig. 7, A, B, D, and E, two-way ANOVA with Tukey’s multiple comparison was performed. One-way ANOVA was performed for Fig. 7 C. A P value <0.05 was considered statistically significant.
Online supplemental material
Fig. S1 shows evolutionary conservation of the missense variants and splicing consequences of the variants affecting slice sites. Fig. S2 shows decreased phosphorylation of MARCKS in the patients’ EBV-B cells. Fig. S3 presents the complementation of the patients’ EBV-B cells with WT PRKCD. Fig. S4 provides additional information about the NADPH oxidase subunit expression and their interaction with PKCδ. Fig. S5 presents additional evidence of impaired ROS production in the patients’ primary cells including ROS production in neutrophils and monocytes, NET formation and phosphorylation of p47phox. Table S1, Table S2, and Table S3 provide primer sequences for sequencing and site-directed mutagenesis, and antibody references.
Acknowledgments
This paper is dedicated to the memory of Prof. Asghar Aghamohammadi, who passed away in November 2020, and made immense contributions to the understanding and treatment of primary immunodeficiencies.
We thank all patients, their relatives, and the treatment teams for their cooperation in this study. We thank all the members of the Laboratory of Human Genetics of Infectious Diseases for helpful discussions. We also thank Christine Rivalain, Cécile Patissier, Lazaro Lorenzo-Diaz, Dominick Papandrea, Dana Liu, and Yelena Nemirovskaya for administrative assistance and Gaspard Kerner for statistical advice. We thank the Necker Institute Imaging Facility for technical advice. The graphical abstract was created with BioRender.com under subscription.
This research was supported by the Yale Center for Mendelian Genomics funded by the National Human Genome Research Institute (UM1HG006504). The Howard Hughes Medical Institute laboratory was funded in part by the National Institute of Allergy and Infectious Diseases (grants 5R01AI089970, 5R37AI095983, and R01AI127564-01), the National Center for Research Resources, and the National Center for Advancing Translational Sciences of the National Institutes of Health (grant 8UL1TR001866 for J-L. Casanova), the Rockefeller University, the St. Giles Foundation, INSERM, the University of Paris, the French Foundation for Medical Research (EQU201903007798), the Integrative Biology of Emerging Infectious Diseases Laboratory of Excellence (ANR-10-LABX-62-IBEID), the SCOR Corporate Foundation for Science, and the French National Research Agency (ANR) under the “Investments for the future” program (grant ANR-10-IAHU-01), ANR-IFNPHOX (grant ANR-13-ISV3-0001-01 for J. Bustamante), ANR-GENMSMD (grant ANR-16-CE17-0005-01 for J. Bustamante), ANR-GENCMCD (grant ANR-11-BSV3–005-01 for A. Puel), and ANR-HGDIFD (grant ANR-14-CE15-0006-01 for J. Bustamante). A-L. Neehus was supported by the Bettencourt Schueller Foundation and the International PhD program of the Imagine Institute, K. Moriya by a Japanese Foundation for Pediatric Research fellowship grant and EURO-CMC project (grant ANR-14-RARE-0005-02 for J. Bustamante), R. Lévy by the Société Nationale Française de Médecine Interne (Bourse Marcel Simon), the INSERM PhD program for medical doctors (poste d’accueil INSERM), and the Fulbright grant (Franco-American commission), T. Le Voyer by the Bettencourt Schueller Foundation and the MD-PhD program of Imagine Institute, and A. Nieto-Patlán by the Consejo Nacional de Ciencia y Tecnología National PhD Fellowship Program. All research at the Great Ormond Street Hospital National Health Service Foundation Trust is supported by the National Institute for Health Research Biomedical Research Centre. P.A. Brogan also acknowledges support from the Great Ormond Street Hospital Charity. N. Lachmann acknowledges support from the Deutsche Forschungsgemeinschaft under Germany's Excellence Strategy - EXC 2155 - project number 390874280 and the REBIRTH Center for Translational Regenerative Medicine funded through the State of Lower Saxony (MWK: ZN3440).
Author contributions: A-L. Neehus, J-L. Casanova, A. Puel, and J. Bustamante conceived and designed the study. A-L. Neehus, K. Moriya, T. Le Voyer, A. Nieto-Patlán, R. Lévy, M. Roynard, K. Haake, M. Migaud, and K. Dorgham performed the experiments. A. Özen, E. Karakoc-Aydiner, S. Baris, A. Yildiran, E. Altundag, F. Dogu, S. Haskologlu, E. İnce, G. Uzel, A. Kiykim, K. Boztug, M.R. Roderick, M. Shahrooei, P.A. Brogan, H. Abolhassani, G. Hancioglu, N. Parvaneh, A. Belot, and A. Ikinciogullari were the treating physicians responsible for clinical diagnosis, deep phenotyping, sample collection, and patient follow-up. G. Gorochov, L. Abel, N. Lachmann, J. El-Benna, and A. Belot provided conceptual advice. A-L. Neehus, J-L. Casanova, A. Puel, and J. Bustamante analyzed the results and wrote the manuscript. All authors revised the manuscript and approved the final manuscript as submitted.
References
Competing Interests
Disclosures: P.A. Brogan reported personal fees from Sobi, Novartis, Roche, and GSK, and grants from Sobi outside the submitted work. No other disclosures were reported.
Author notes
K. Moriya and A. Nieto-Patlán contributed equally to this paper.
F. Dogu, S. Haskologlu, E. İnce, J. El-Benna, G. Uzel, A. Kiykim, K. Boztug, M.R. Roderick, M. Shahrooei, P.A. Brogan, H. Abolhassani, G. Hancioglu, N. Parvaneh, A. Belot, and A. Ikinciogullari contributed equally to this paper.
J-L. Casanova, A. Puel, and J. Bustamante contributed equally to this paper.
N. Lachmann's present address is Department for Pediatric Pneumology, Allergology and Neonatology, Hannover Medical School, RESIST Cluster of Excellence, Hannover, Germany.